fastglm 0.1.4

• Sparse designs are now read with Eigen::Map<Eigen::SparseMatrix> instead of Eigen::MappedSparseMatrix, which Eigen 5 removes, so the package builds with both the current and the upcoming ‘RcppEigen’.
• Constrained cumulative-link fits with no predictor columns no longer form a row block of an empty matrix, which offset a null pointer.
• The count model kernels no longer call realloc() with size 0 when a data block has no rows with positive weight.
• The Parquet source test is skipped when ‘arrow’ is built without Parquet support.

fastglm 0.1.3

Behavior changes

• Requires R 4.1.0 or later.
• Aliased coefficients from rank-deficient designs are reported as NA, as in glm(), instead of 0.
• When the Cholesky factorization of a dense design is singular or ill-conditioned, method = 2 and method = 3 now refit with pivoted QR and warn, instead of returning arbitrary coefficients or invalid covariance. Sparse and big.matrix designs, which cannot use QR, report an error.
• Indefinite mixture information no longer produces standard errors from absolute values, and information matrices are evaluated at the reported estimates.
• Streaming fits now validate every chunk and propagate source errors.
• Printed output for models, summaries, diagnostics, predictions, bootstraps, and tests now shares one format. Numeric fields are unchanged, and test results and predictions support as.data.frame().
• The plot method for ordinal assessments uses ggplot2 when it is installed.

Ordinal models

• Added fastglm_ordinal() for cumulative-link models with logit, probit, and complementary log-log links, by maximum likelihood or mean bias reduction, with formula and matrix interfaces, frequency counts, and offsets.
• Added threshold-specific effects through nonparallel, including partial proportional odds. Fits whose best solution lies on the ordering boundary stop with convergence.reason = "ordering_boundary" and an explanation.
• Added category and cumulative predictions with standard errors, average probabilities and contrasts, fastglm_marginal_effects(), simulation, residuals, and optional broom methods.
• Added fastglm_ordinal_assess() for prediction scores and calibration, fastglm_po_test() for common-slope tests, and boundary-row diagnostics through fastglm_diagnostics(boundary = TRUE).
• Added robust and clustered covariance, refitting bootstraps for common-slope fits, profile-likelihood intervals, and score tests.
• Added fastglm_ordinal_prepare() and fastglm_ordinal_batch() for repeated responses, and fastglm_ordinal_streaming() for data read in blocks.

GLMs and count models

• Added a formula interface, named and automatic solver choices, prediction with offsets, leverage, simulation, Wald intervals, numerical diagnostics, and optional broom methods.
• Added fastglm_prepare() and fastglm_batch() for repeated outcomes on a common design, including bias reduction, optional process workers, and inference = "none" to skip covariance.
• Added fastglm_truncated() for zero-truncated Poisson and negative-binomial regression, and fastglm_twopart() for nonnegative continuous outcomes.
• Count models now provide component and marginal predictions, probability masses, delta-method standard errors, residuals, simulation, and profile intervals for theta. Negative-binomial, hurdle, and two-part fits also accept sparse designs and repeatable sources.
• Added joint robust and clustered covariance for count and two-part models, average response contrasts, and refitting bootstraps.

Inference

• Added fastglm_wald_test() for joint term tests and named contrasts in GLMs, ordinal, truncated, hurdle, zero-inflated, and two-part models.
• Added profile-likelihood intervals and fastglm_profile_test() for dense binomial-logit Firth fits.

Large data

• Added method = "qr" to streaming GLMs, including mean bias reduction.
• Added resettable sources with fixed formula schemas, optional DBI and Parquet adapters, sinks for predictions and diagnostics, robust and clustered covariance, and average predictions for streamed fits.

Numerical corrections

• Binomial probit and complementary log-log bias reduction now evaluates tail probabilities, scores, and information accurately for nearly deterministic outcomes, and limits oversized scoring updates.
• Gamma mean bias reduction now solves the adjusted dispersion equation and guards against overshooting initial steps. Dense Firth fits with a zero-rank working design no longer report convergence.
• Negative-binomial theta scores and likelihoods are stable near the Poisson limit. Boundary theta estimates have unavailable standard errors, and hurdle negative-binomial covariance uses joint observed information.
• Corrected rank-deficient QR updates, aliased coefficient reporting, mapped matrix views, offset null models, zero-weight degrees of freedom, grouped-binomial likelihoods, and simulation of numeric binomial proportions.
• fastglm_ordinal() now stops when named count-matrix columns do not match levels in order. These columns were previously assigned by position, which could silently reverse a fit.
• Printed tables without p-values no longer round small estimates, standard errors, or bootstrap limits to zero, and models without a size parameter no longer print a missing Theta row.
• Streamed fits no longer retain the last source block, so saving a fit does not write those rows.

Performance

• Ordinal maximum likelihood begins with observed-information Newton steps, whose cost does not grow with the number of categories. Expected information still certifies convergence and gives the reported covariance. control = list(observed = FALSE) restores expected scoring throughout.
• Ordinal fits with many categories use a block factorization, and setup, robust covariance, common-slope tests, and constrained fits run in C++.
• Negative-binomial, hurdle, and truncated-count calculations for dense designs run in C++, and Firth fits reuse constant matrix calculations where possible.

fastglm 0.1.2

Bug fixes

• Compilation now works under libc++ with clang 23 and newer, which no longer supplies <iterator> and <algorithm> as transitive includes. The bigmemory header BigMatrix.h, pulled in by the big.matrix backend, uses std::back_inserter and std::copy without including those headers itself, so the two source files that include it (fit_glm_dense.cpp and bigmemory.cpp) now include <iterator> and <algorithm> explicitly beforehand. Reported by CRAN’s clang-trunk checks.

fastglm 0.1.1

New features

• Firth bias-reduced GLMs (firth = TRUE) now work for all standard families (gaussian, binomial, Poisson, Gamma, inverse.gaussian) on dense, sparse, and streaming backends. In 0.1.0 Firth was limited to binomial(link = "logit") on dense designs only.

• SQUAREM acceleration (Varadhan and Roland, 2008) for the fastglm_zi() EM driver, converting linear EM convergence to near-quadratic.

Numerical stability

• Negative-binomial initialization in fastglm_nb() now uses a moment-based mu/theta seed instead of always running a full pilot Poisson fit, improving convergence on overdispersed data and near boundary cases.

• Increased the default outer.maxit for fastglm_nb() from 25 to 50 to allow convergence in harder NB problems.

• Clamping guards in native C++ family kernels for inverse-link and sqrt-link families (Gamma, inverse.gaussian, Tweedie) to prevent overflow when eta is near zero.

• Floor on mu.eta for Tweedie inverse and sqrt links to avoid division by zero in IRLS weights.

• Sign correction in the Tweedie sqrt mu.eta kernel.

Bug fixes

• Replaced Rf_error() with Rcpp::stop() in bigmemory.cpp (reported in GitHub issue).

Documentation

• New vignette firth-fastglm covering Firth bias-reduced GLMs for all supported families.

• New vignettes nb-convergence-fastglm and nb-stability-fastglm demonstrating fastglm_nb() convergence and stability relative to MASS::glm.nb() on challenging datasets.

• Expanded benchmarks vignette with additional model classes.

Internal

• Shared Brent root-finder moved to inst/include/brent.h, replacing three duplicated implementations.

• Expanded test suite (107 test blocks / 361 expectations, up from 54 / 131 in 0.1.0), including new suites for numerical edge cases and Firth across all families.

fastglm 0.1.0

Breaking changes

• fastglm.fit() and fastglm.control() have been renamed to fastglm_fit() and fastglm_control(). The old dot-form names triggered an apparent-S3-method R CMD check NOTE because fastglm is itself a generic. Existing code that passes method = fastglm.fit to glm() should be updated to method = fastglm_fit.

New features

• New top-level function fastglm_nb() for negative-binomial regression with the dispersion theta estimated jointly with the regression coefficients. Plays the same role as MASS::glm.nb(), but the IRLS loop, the inner Brent root-find for theta, and the outer (beta, theta) alternation all run entirely in C++.

• New top-level function fastglm_hurdle() for two-part count models with a binary zero / non-zero component and a zero-truncated Poisson or NB count component. Same model as pscl::hurdle(); the joint fit runs in a single C++ driver that calls the existing IRLS solver twice and runs an inner Brent MLE for theta in the unknown-theta NB case.

• New top-level function fastglm_zi() for zero-inflated Poisson and NB regression. Same model as pscl::zeroinfl(); the entire EM driver runs in C++, including closed-form posterior responsibilities, weighted IRLS for both M-steps, the inner Brent MLE for theta, and the analytical observed-information vcov.

• New firth = TRUE argument to fastglm() and fastglm_fit() for Firth’s bias-reducing penalty on the score. Currently supported for family = binomial(link = "logit") on dense designs; converges in finite steps under separation, where unpenalized IRLS would diverge. Coefficients agree with logistf::logistf() to 1e-7 on the standard Heinze-Schemper test cases.

• New built-in negbin(theta, link) family for negative-binomial regression with known dispersion, dispatched to a native C++ kernel on the log, sqrt, and identity links. Drop-in for MASS::negative.binomial(theta, link) without a hard dependency on MASS.

• Native fast paths for the family$family == "Negative Binomial(K)" form produced by MASS::negative.binomial(), the quasibinomial() / quasipoisson() families, and statmod::tweedie(). Detection is automatic.

Documentation

• New vignette count-firth-fastglm covering fastglm_nb(), fastglm_hurdle(), fastglm_zi(), and the firth = TRUE flag, with small inline accuracy and timing comparisons against MASS::glm.nb, pscl::hurdle, pscl::zeroinfl, and logistf::logistf.

• New vignette benchmarks-fastglm providing a more comprehensive benchmark study at larger sample sizes, covering all six model classes. Pre-compiled at maintainer build time so it does not run during R CMD check / R CMD build and adds essentially zero time to the package check budget.

• fastglm-overview updated with a short section introducing the new count-data and Firth entry points.

Other new features

• Real vcov() for "fastglm" objects: the unscaled covariance is now computed in C++ from the IRLS factorisation directly and exposed as $cov.unscaled. summary.fastglm(), vcov.fastglm(), and predict.fastglm(se.fit = TRUE) now all work without a refit hack; vcov.fastglmFit() no longer re-runs glm.fit to recover a qr slot.

• S3 methods registered on sandwich::vcovHC() and sandwich::vcovCL() for heteroskedasticity-consistent (HC0–HC3) and cluster-robust covariance matrices. After library(sandwich), vcovHC(fit) and vcovCL(fit, ...) match sandwich::vcovHC.glm() / sandwich::vcovCL.glm() to floating-point precision and work for sparse, big.matrix, and in-memory fits. (Earlier development versions shipped local vcovHC() / vcovCL() generics; those have been removed in favor of registering on the canonical sandwich generics.)

• Sparse design matrices (Matrix::dgCMatrix) are now supported directly by fastglm() and fastglm_fit() for the LLT (method = 2) and LDLT (method = 3) Cholesky paths. Other decompositions are rejected with an informative error rather than silently densified.

• bigmemory::big.matrix inputs now stream the design matrix in row-blocks of FASTGLM_CHUNK_ROWS rows (default 16384, user-configurable via the environment variable). Filebacked big.matrix objects no longer have to be materialised in RAM.

• New top-level function fastglm_streaming(chunk_callback, n_chunks, family, ...) for fitting GLMs on data sources that do not fit in memory: Arrow datasets, Parquet files, DuckDB queries, CSV streamers, and any other chunk-yielding iterator. The IRLS loop, step-halving, and Cholesky solve all run in C++; the closure is invoked only to deliver one row-block at a time.

Speed

• For standard families (gaussian / binomial / poisson / Gamma / inverse.gaussian on their common links) the per-iteration calls to family$variance, mu.eta, linkinv, and dev.resids are now evaluated in inline C++ rather than via R callbacks. Detection is automatic; non-standard families fall back to the previous R-callback path with no user-facing change.

• The IRLS solver pre-allocates its working buffers across iterations and uses noalias() writes throughout solve_wls(). Eigen’s parallelism is no longer disabled.

• On large n the combined effect is roughly a 1.5×–2× speed-up over fastglm 0.0.4 on the same hardware, on top of the existing 3×–10× advantage over stats::glm().

Documentation

• New vignette fastglm-overview providing a single high-level entry point covering all of the package’s functionality.

• New vignette large-data-fastglm walking through the three large-data paths (sparse, big.matrix, streaming callback) end to end, including the Arrow / Parquet recipe.

Internal

• New tests/testthat/ suite covering vcov(), native vs callback family dispatch, sparse fits, big.matrix streaming, callback-based streaming, and robust SE.

fastglm 0.0.4 and earlier

• Added a NEWS.md file to track changes to the package.

• Added fastglm.fit() to be used as a fitting method for glm().

• Documentation and vignette updates.